Just where do you look for that final 20 percent that puts your mix on a par with commercial tracks? Find out, as we apply a little polish to Zeno's track 'Signs'.
Like many tracks that are submitted to Mix Rescue, this month's mix already seemed to be about 80 percent right. The song was well put together and confidently performed, with respectable tracking quality and appropriate changes in instrumentation between sections to clarify the structure. There were also a lot of sensible production enhancements: well‑chosen snare and kick samples had been triggered alongside the live kit; some effective multi‑tracked claps had been added; and the guitar sound had been built up from multiple layers, in keeping with a wide‑screen style reminiscent of bands like Muse and Biffy Clyro.
Despite all this goodness, when I heard the band's mix for the first time on the SOS Forum, I couldn't help feeling slightly underwhelmed. While nothing was very obviously 'wrong', many parts (perhaps the drums most of all) came across as a bit limp and distant, while the whole package was unnecessarily clouded, without sounding 'beefy'. It can be difficult to pin down what needs doing in such cases, so I asked the lead singer and guitarist, Sam Houghton, if he'd send me the multitrack files, so that I could diagnose the problems and maybe push things a bit further.
There turned out to be a number of different areas of the production where I felt the band might unearth that elusive 20 percent — and as these areas are also make‑or‑break issues for many other self‑produced mixes I've heard, I'll focus this month on how I addressed them.
Zeno's drum recordings were already pretty good, and because they'd hedged their bets by recording alternative kick and snare mics and including good triggered samples, little needed to be done with EQ, other than cutting out low‑end clutter with high‑pass filtering and giving those polarity‑inversion buttons a good flex. Much more of an issue, to my ears, was that this kind of song was crying out for obviously compressed‑sounding drums. There's a time and a place for natural drum dynamics, but this wasn't it: the song called for something much more bombastic and unruly. This was one of the key areas where Zeno's original mix fell short, because they'd made the common mistake of relying too much on reverb to create size and sustain, which had allowed the drums to become too slender and distant.
To that end, my first port of call was the overhead mics, the plan being to pulverise them with some kind of 'character' compressor. I decided to use the opportunity to try out the new Empirical Labs' EL7 Fatso Snr plug‑in on Universal Audio's UAD2 platform. To start with, I turned up the input‑level control and experimented with the Comp (compressor type), Attack, and Release switches, in search of that compression effect where everything seems to be fighting everything else, and where the cymbals sustain almost indefinitely. This typically requires lots of medium‑ratio gain reduction with fast attack and release times, but there's something of an art in finding the exact settings, because too high a ratio or too short an attack can flatten the drum peaks too much, while too short a release reduces the rowdy dynamic side‑effects and scarcely ruffles the listener's cravat.
This kind of processing tends to work best when there are also room mics to blend in with the overheads, but this was one option Zeno hadn't planned for while tracking, so I synthesized something akin to this with a separate short reverb (an impulse response in Christian Knufinke's SIR2, simulating roomier overhead mics), which was high‑pass filtered at 230Hz and bussed, along with the overheads, to the Fatso. As usual, I tried a few different likely‑looking impulse responses before finding one that seemed to fit, and then I tweaked it further to improve its suitability — by simply reducing the length to achieve a decay time of just under a second. Once that was all sorted, I turned back to the Fatso and aimed for bonus points with the more colourful Warmth and Tranny facilities, which repaid some 'suck it and see' twiddling with a useful extra thickening of the sound.
Emboldened by my overheads processing, I threw caution to the winds where the snare and tom‑tom fills tracks were concerned. In the former case, I fattened the sound by squeezing it through SSL's famously savage Listen Mic Compressor, while the tom‑toms were mashed through Audio Damage's rapacious Rough Rider compressor with fast attack and release settings, to create a kind of explosive sound that seemed to make a nice statement whenever it occurred. I did have to be careful to clean up some odd bits of snare and cymbal spill on this track, though, because they quickly made a nonsense of the drum balance when shot into orbit by Rough Rider's make‑up gain between tom hits.
The individual track processing was only part of the equation, though, because I also bussed all of the drums to an instance of Universal Audio's SSL‑inspired 4k Buss Compressor, in order to create a more obvious rhythmic pumping effect from the kick and snare hits. Again, the principles of setting up this kind of effect are pretty simple: configure the compressor to generate a few decibels of gain reduction on each kick and snare hit, and then adjust the release time for the nicest‑sounding 'sucky' gain‑riding. But the devil's in the detail...
A common mistake is to balance the kick and snare too low in the drum mix, so the compressor doesn't have large enough peaks to react to. These tracks can afford to be good and loud if you're after an obvious pumping effect, because the compressor will, of course, be doing its best to clobber them back into the drum mix. Naturally, the release setting is critical to the actual pumping effect itself, but the attack time is no less important, for two reasons: firstly, it'll determine how much gain reduction (and therefore pumping) you get on each instrument for a given compressor threshold setting; and, secondly, it can drastically alter the important attack transients of these instruments. Set the attack too slow and it's trickier to get dramatic pumping; set it too fast and your kick and snare transients begin to lose both snap and low‑end weight.
This mix was a case in point — so much so, in fact, that I couldn't find a compromise I was happy with in this respect. It wasn't that I didn't like the SSL‑style compressor pumping — I did — but that, given the amount I was compressing (up to 8dB), even quite a slow 10ms attack was blunting the main transients too much for my taste. I could have tried a few other compressors here, but because I mostly liked what I was hearing, I decided to take a different tack, using Reaper's newly introduced per‑plug‑in wet/dry mix control to feed a little unprocessed signal past the compressor. However, by the time the snare sounded the way I wanted it to, the kick was still being robbed of attack and low end, so I used one of the kick‑drum channel's sends to route further unprocessed signal direct to the main mix buss. Finally, a result!
Further down the line, I also compressed the main mix buss, using one of the faster auto‑release Fairchild 670 emulations in URS's Console Strip Pro, a good general‑purpose plug‑in I often have strapped across my mix for buss‑processing purposes. Why the Fairchild 670? For no other reason than that I tried a half a dozen different emulations and in the end that one seemed to provide the best compromise between pumping and transient definition, just as before. I played safer with this compressor than with the 4k Buss Compressor, though, allowing a maximum of 4dB of reduction on the loudest peaks. Nevertheless, all in all I'd clearly dialled in enough compression to make anyone of a sensitive disposition swoon — but I want to stress that the purpose of this wasn't to iron out the drum peaks for the sake of perceived loudness: if that's your aim, it's best left to specialised processing at the mastering stage. Even with all the mix compression I used, the drum hits were still clearly visible as big spikes sticking out of the exported mixdown's waveform plot. Only later, once the mix was basically finished, did I lprocess the bounced file with mastering‑style saturation and limiting to bring its loudness into line with the kind of bands Zeno were competing with. Loudness maximisation has to be finely balanced to minimise its unwanted side‑effects, and trying to do it while mixing is almost always a recipe for disaster.
As I demonstrated in SOS December 2009 with Ollie Wright's song 'All The Same' (/sos/dec09/articles/mixrescue_1209.htm), ensuring that the timing of rhythm instruments and lead vocals corresponds to that of the drums can greatly improve an arrangement's sense of drive, and this applied equally here. Although Zeno's drum parts were providing a solid backbone for the groove, the bass part and a number of the guitars weren't supporting this enough, and the lead vocals (which would presumably have been tracked against the bass and guitars) were another culprit in this department too. It wasn't tricky to tighten these up with my sequencer's audio editing tools, and the improvements to the punchiness of the rhythm were more than adequate compensation for the tedium of the editing job!
Another problem that 'Signs' shared with 'All The Same' was that, in general, the guitars had been overdriven more than necessary, and this was responsible for a certain amount of 'mushiness' during the choruses. There was a cleaner part amongst those that Sam had provided, so part of my solution was to shift the mix balance away from the grungier parts to obtain more pitch definition — but that left the overall guitar sound rather 'pillowy'. Fortunately, Sam had recorded guitar DI feeds for each part, so I was able to perk things up by re-amping the main double‑tracked part through the Fender Bassman model from the Jimi Hendrix edition of IK Multimedia's Amplitube. I kept the preamp's gain around midway to avoid too much fuzz, turned the bass EQ knob down to improve the mid‑range cut‑through, and pushed the volume control right up to add some extra attitude from the cabinet model. I find that it can help opposition‑panned double‑tracked guitars to sound bigger if the two sides don't sound exactly the same, so as a final touch I swapped out the Bassman's EQ model on one channel for that of a Fender Dual Showman.
Most of the other sonic questions that arose in Zeno's 'Signs' I answered simply by building up the track in order of importance, starting with the final chorus and then putting together the remaining song sections one by one. I also started the chorus sections of the mix with the bass sound, constructing the guitar arrangement around it, primarily so that I didn't have to rely too heavily on the panned guitars in the balance and could therefore leave a little more space at the left and right sides of the mix for the synths and piano, which were all but submerged in the original mix. As such, Zeno could probably have improved their own mix simply by taking a more structured approach.
However, I did call on a couple of specific tricks that go beyond the general principles of mixing technique. The first concerned the lead vocals in the choruses, which were sounding quite muffled against the distorted bass/guitar‑led wall of sound. The proximity‑effect bass boost of the tracking mic was a big part of the problem here, but even with a stiff low‑end EQ roll‑off addressing this, the vocals were still heavily masked in the 3kHz presence region by the distorted instruments, and therefore sounded soft and recessed. In Zeno's original mix, the band's response had been to fade the vocal up higher, and while this helped unmask it, it also made the rest of the arrangement feel smaller by comparison.
It's tempting to over‑EQ in this situation, either by boosting heavily around 3kHz on the vocal — which almost always leads to a 'telephonic' harshness before it cures the problem — or by cutting too much from the guitars at the very frequencies that give them their grit. For this reason, many engineers choose to add distortion components to the vocal sound instead, to level the playing field, and that was the angle I adopted, turning to an emulation of the Pro Co Rat distortion stomp‑box (a popular choice), courtesy of IK Multimedia's Amplitube Metal processing bundle. As with any distortion effect you might add at mixdown, it's worth sculpting its output carefully with EQ to focus it into the frequency range required, and in my case I did this with a general‑purpose graphic-EQ stomp box chained after the Rat emulation.
The other little trick that was particularly handy was that of listening back to the balance during the final stages of the mix with the most prominent instruments bypassed. Here, that meant muting various combinations of the drums, bass, and lead vocals, and the reason why it's so useful is that it focuses your attention on more subtle background balances that might not be as clearly apparent otherwise. It proved particularly useful for evaluating the levels and frequency ranges of the send effects during the choruses, helping to keep the mix as uncluttered and up-front as possible, but it's also great for confirming the exact fader levels of background instruments and pads.
As you can hear if you listen to the 'before' and 'after' example files this month (see the box, above right), there was more to this mix than processing and effects work. This is because it struck me that there didn't seem to be enough arrangement interest during the verses and middle section to sustain the listener's attention (rarely a good thing in any commercial style). Elsewhere, there were specific tracks that seemed a little too repetitive, as well: the guitar riff first heard in the introduction, for example, or the filter‑modulated synth pad featured prominently in the two verses. Also, both verses remained quite static in terms of instrumentation, with very little to bridge the quite long gaps between vocal phrases or build the arrangement in the shorter term. Sam had said to me at the outset that the band were open to new input, so I decided to have a crack at some more creative changes.
It usually makes sense to take a cue from the existing material, and the first thing I ran with was the idea of using delay effects to generate some interest between the vocal phrases — which is something that Zeno had already explored to some extent in their mix. Like them, I introduced a tempo‑sync'ed delay, but I decided to halve the delay time (for a faster rate of repeats) and then ride the delay in and out more frequently. However, I combined this with an extremely long special‑effect reverb from one of my favourite bits of freeware, Da Sample's Glaceverb. I used Voxengo's MSED to reduce the stereo width a little (because by default it seemed rather all‑encompassing) and then filtered the output for a mid-range‑heavy tone distinct from the vocal itself. Unlike the delay, however, I fed the reverb effect from a separate duplicate lead‑vocal track, so that I could easily edit the audio to vary which notes had reverb, and when the onset of the reverb occurred in each case. The reason I did this was that I wanted to undermine any expectation that the effect would remain constant, and thereby retain the listener's interest for longer. I was also able to time‑stretch the last note of both verses to create a neat transition into the pounding pre-chorus sections.
These send effects helped a great deal, but it's tough to keep any treatment from going stale over the course of a whole 16‑bar verse, so I immediately did a couple of edits to mute elements from the first half of each verse, thereby creating some arrangement build‑up: in the first verse, I nixed the first eight bars of synth pad, while in the second verse I ditched corresponding sections of the multitracked claps and effected piano. In tandem with this, I also worked in some fills for each verse's two eight‑bar section boundaries. The electric guitar riff from the first verse (with a sprinkling of Reaper's Elastique 2.1 Pro time‑stretching) provides the basis for a fill midway through the verse, while the long vocal reverb tail and some drum edits generated the fill that leads into the pre‑chorus. The second verse, on the other hand, uses the multitracked claps and some guitar volume swells to achieve similar ends.
Another aspect of the track that I took as inspiration for some additions was the main rhythmic synth pad. Although this added a nice pinch of electronica to the brew, it did end up being a bit too repetitive for me. My course of action was to bring in a couple of rhythmic synth loops from Nine Volt Audio's Melodic REX library (one of my favourite titles from their excellent BPM Flex series), and then sync and tune those to the track using more Elastique processing. This is such a simple thing to do and can really elevate the variety and detail of a mix, even if (as in this case) you keep the loops mostly in the background of the mix to avoid changing the song's overall character too much.
The transformation of the middle section was little more than a combination of similar techniques, although I did also spread the delay effects on with a trowel and mangle the piano sound through Schwa's Oligarc, Betabugs' Flofi and MDA's Leslie.
Another common shortcoming of Zeno's original mix was that they'd not really taken advantage of the power of automation to maintain the most effective balance. I'm always harping on about this, especially with regard to lead vocal levels, and I make no apologies for that — you've just got to roll up your sleeves and get busy with fader rides if you're after commercial‑sounding results in almost any style these days. I've dealt with specific automation techniques many times in the past (in the SOS August, October, and November 2009 columns, for example), so I won't retread that ground here, but I will add that fine balance issues can be very difficult to judge on project‑studio stereo monitoring systems such as headphones and nearfield monitors.
The problem is that all the important centrally panned sounds in your mix are only heard as an illusion in stereo, what is often called a 'phantom image', because there's no actual centre speaker. The vagaries of speaker design and room acoustics mean that phantom images can be rather indistinct and unstable in real‑world small‑studio monitoring environments, which makes critical balance judgements tricky. If you listen to your mix in mono from a single speaker, however, you have no phantom image, which means that balance problems normally stand out much more clearly. And while we're on the subject of monitoring technique, I also suspect that Zeno (like a lot of other recording musicians) need to take more time to compare their material against competing commercial tracks. This makes many balance decisions much more objective, and also gives a valuable reality check on the overall tonality of the mix.
If your mix feels as though it's only a step away from a polished sound, but you're not sure what that step is, the chances are that it's actually a combination of different mini‑steps that is required. This month, I've demonstrated what this meant in the case of Zeno's 'Signs', but because the same afflictions so commonly affect budget productions, similar remedies may be just the ticket for your mix too.
We've placed before and after sound files in both MP3 and WAV format at /sos/mar10/articles/mixrescueaudio.htm so you can hear the impact of the techniques described in this article.
Leeds-based three‑piece alt‑rock band Zeno cite The Killers, Feeder, Muse, Biffy Clyro, Radiohead, and Sigur Ros amongst their influences. The band includes Sam Houghton (above) on guitar and vocals, Josh Hart (bass, vocals), and Tim Gosden (drums, synths), with Sam and Josh co‑writing the song featured in this month's column. The band have been going since late 2009, are on the lookout for gigs, and are working on an upcoming four‑track EP, Transhuman.
Zeno: "Wow! Crazy. Different. The first thing we all noticed when listening to Mike's mix was how cool the vocals sounded — he's done a great job, especially considering how they were recorded. One thing we all love is how the delay rises and spreads out of the mix. It makes the verses much more interesting and atmospheric. The guitar tones also work really well for the song, and the kick and bass sit nicely together. The overall character of the track has changed for the better, too. It's moved from rock to a much more electronic vibe, something we are excited about incorporating in our live sets. We're definitely going to be using some of Mike's techniques on our forthcoming EP. Many thanks to Mike and Sound On Sound!”